-
Notifications
You must be signed in to change notification settings - Fork 114
Process crashes when using two NLE instances sequentially (on MacOS for Debug Builds). #254
Comments
This appears to only trigger on my personal machine, not on CI or for anyone else. Closing for now. |
OK, this does break on CI as well, but only (1) on MacOS, and (2) when using a Debug build: https://github.com/facebookresearch/nle/runs/4359406818?check_suite_focus=true |
Issue demonstrated in #290. |
Fail if libnethack is resident before dlopening. This is the issue in #254
Related to the dlopen/dlclose dance not actually closing in this specific case (which it never guaranteed to do), as in this issue. Possible solution: https://gist.github.com/heiner/bc78064fec32174e1a216dbd5fbc6503 |
Based on investigation in https://gist.github.com/heiner/bc78064fec32174e1a216dbd5fbc6503 Fixes #254.
Hi, just wondering if it crashes with an error message at all. I'm getting a EDIT: I got it working by reducing the action space as I noticed it was specifically happening when executing specific actions. |
🐛 Bug
The test in #253 should pass but fails on MacOS for Debug builds.
To Reproduce
Environment
Collecting environment information...
NLE version: 0.7.3+08b9280
PyTorch version: 1.9.0
Is debug build: No
CUDA used to build PyTorch: None
OS: Mac OSX 11.5.1
GCC version: Could not collect
CMake version: version 3.20.0
Python version: 3.8
Is CUDA available: No
CUDA runtime version: No CUDA
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Versions of relevant libraries:
[pip3] numpy==1.19.2
[pip3] numpysane==0.34
[pip3] torch==1.9.0
[conda] blas 1.0 mkl
[conda] mkl 2019.4 233
[conda] mkl-service 2.3.0 py38h9ed2024_0
[conda] mkl_fft 1.3.0 py38ha059aab_0
[conda] mkl_random 1.1.1 py38h959d312_0
[conda] pytorch 1.9.0 py3.8_0 pytorch
The text was updated successfully, but these errors were encountered: